Intel's breakthrough in incorporating the above mentioned new materials to progress transistor technology was first announced in 2003 and was successfully put to the test in manufacturing a 45nm SRAM part in 2006. Not long back, they even have got the first Penryn silicon fully up and running. As we probed Intel further on the status of Penryn at the moment, they commented that the working silicon as of now is far from its final clock rates and is currently operating well below the speeds of a typical Core 2 Duo. However, they expect to see mature yields of the Penryn silicon in the second half of 2007 and they are on track to begin volume production in that time frame with first retail products to hit somewhere in first half of 2008. No specific dates were made known yet for the actual product launch as it is subjected to the progress of their platform vendors and other partners.
On the note of platforms and the likes, Intel could not yet comment of Penryn processor's compatibility with existing Core 2 Duo motherboards. While still using the LGA775 socket, it is still too early to determine the final compatibility of a product that's a year away. The best outcome is of course a BIOS flash to existing boards, but we have a hunch that Intel will draft yet another power spec for the processor, which may be incompatible with current boards or even require another chipset for proper compatibility. Intel has mentioned that Penryn will support new power management modes and that alone could be enough cause for an updated platform as well.
We have to concede that Intel is really using its skill in semiconductor technology in both design and manufacturing prowess and more so when you look at Intel's committed goals. Their hard hitting penetration of the Core 2 Duo, Core 2 Quad products and derivatives for the server space have been swift and effective. Successful internal demo of the 45nm process based processor and two 300mm wafer factories committed to manufacturing these 45nm products and a third fabrication plant by first half of 2008 are no small achievements. With the Spring edition of Intel Developer Forum 2007 just round the corner, you can bet that Penryn would be showcased and media like us will be there to give you a first-hand glimpse too. Not only these achievements, but looking back at its past lithography process technology and timelines, they have continuously matched a 2-year cycle of advancement. Thus by 2009 and 2011, Intel expects to see 32nm and 22nm process technology production runs to commence respectively.
As rosy as all these sound, Intel's path is far from being clear as AMD's next generation microarchitecture processor family named Barcelona for the server space and Agena for the enthusiast segment are expected to spring in to action just after the mid of this year. Touting a true native quad-core architecture, this will easily allow AMD to scale to a platform with eight or even 16 processing cores (in the form of dual or quad physical processors) based on existing server designs. Not only is AMD squeezing four cores in one die based on the 65nm process technology, but each of these cores have been overhauled as well with better floating point prowess and memory tweaks to give it a stronger comeback punch - if all their plans go well. AMD expects to catch up with Intel in the manufacturing process technology race somewhere in 2008 as well in the most optimistic scenario, but of course they would first have to prove their worth by delivering their upcoming true quad-core chips later this year as promised.
While Intel is expected to launch much faster Core 2 Quad variants later this year and with new platform technologies, it's still early days to figure out who's playing the right cards. Intel may probably safely ride out the first half of this year, but only in the later half can we truly judge who has got the upper hand for the entire 2007.